[HN Gopher] How to Blow Up a Timeline
___________________________________________________________________
How to Blow Up a Timeline
Author : jger15
Score : 62 points
Date : 2023-07-08 09:57 UTC (13 hours ago)
(HTM) web link (www.eugenewei.com)
(TXT) w3m dump (www.eugenewei.com)
| PaulHoule wrote:
| Personally I am interested in the comparisons to TikTok that is a
| subject people don't really understand.
|
| One of the worst problems of many recommendation systems is that
| they don't get definite feedback about what you hate, only about
| what you like. They have to guess about negative examples but
| it's just a guess. Systems like YouTube that show you a large
| number of links that might attract you look smarter in the short
| term because they can ride on your ability to choose something,
| but they can draw very little conclusion about the links you
| didn't click on.
|
| People forgot that StumbleUpon was a recommender for web sites
| that had an interface very similar to TikTok, my YOShInOn smart
| RSS reader also works like TikTok but it is a pure content-based
| recommender.
|
| People say that "engagement=outrage", something that is only
| about 80% true (my predictive model believes that HN participants
| like to talk about cars.). In a network like Twitter or Mastodon
| where people share posts the classic social feed very much
| amplifies outrage.
|
| If I can actually bear to look at 5000 angry toots I might try
| training an anti-outrage model. This could be used several ways:
| (1) at the input of a reader (I just personally don't want to see
| it), (2) at the output of the write (I know, and I get the
| feedback, that YOShInOn is a nicer "person" than I am) or (3) at
| the heart of a network to suppress the cycle of anger
| amplification. You don't even need to get all of it, you just
| have to reduce the coefficient of transmission as you would for a
| virus to improve the tone of a a space.
|
| (3) would by far be the most effective but boy would it get
| pushback from people, whereas I can apply (1) and (2) to myself
| and nobody knows.
|
| Look closely at an "angry" person and you will find they are
| "angry about politics" or "gay and angry" or "conservative and
| angry" and a model like that really will suppress "politics" and
| "gay" and "conservative" (the conditional probability
| distribution says to do so) and more so will be accused of doing
| so.
|
| The argument that almost moves me is that networks like that have
| the potential to enable organizing for change and that
| suppressing the visibility of outrage would squelch that but we
| are already seeing outrage-driven politics reach a point where
| people choose their values to fit their tribe instead of choosing
| a tribe that fits their values. More and more people will only
| vote for one party ever which leads to a "doom loop" in politics
| because there is no competition for "effective", "not corrupt",
| "get what you pay for with your taxes" which are in (almost)
| everyone's interest. Trouble is that is the road to Lebanon, at
| best you have corrupt politicians stealing everything and a
| government that has no legitimacy because it can't get anything
| done. At worst people will tear it all down.
| KineticLensman wrote:
| > They have to guess about negative examples but it's just a
| guess.
|
| For those who don't use TikTok, there are three negative
| signals that you can send.
|
| * (implicit) whenever you scroll to the next clip before the
| current one has finished
|
| * long pressing and selecting 'not interested'. When you do
| this to an ad, you have to select a reason.
|
| * reporting (flagging) a clip for a specified transgression.
| refulgentis wrote:
| Lost a significant chunk of my attention with the thesis, then
| all of it in a thicket of tendacious musings.
|
| > "Twitter will persist in a deteriorated state, perhaps
| indefinitely."
|
| Makes sense.
|
| > "However, it's already a pale shadow of what it was at its
| peak. The cool kids are no longer sitting over in bottle service
| knocking out banger tweets."
|
| Fun visually evocative analogy, but gossamer. As soon as we try
| applying it to reality, it falls apart.
|
| > "Instead, the timeline is filled with more and more strangers
| the bouncer let in to shill their tweetstorms, many of them
| Twitter Verified accounts who paid the grand fee of $8 a month
| for the privilege."
|
| What's "the timeline" here? For You? Following? This hides the
| ball on the fact there's a much simpler concept here that we're
| already familiar with, algorithmic vs. "true" timeline. In fact,
| Twitter pre-Musk had this. It's just different bluechecks now.
|
| > "In the past year, so many random meetings I have with one-time
| Twitter junkies begin with a long sigh and then a question that
| is more lamentation than anything else: "How did Twitter get so
| bad?""
|
| From above: the cool kids aren't in bottle service banging
| (whatever that means) and they're stuck on the algorithmic
| timeline.
| PaulHoule wrote:
| The article was too long but I thought had some interesting
| ideas.
|
| Lately though I think the key to understanding Twitter is the
| relationship with the media and particularly "media
| amplification"
|
| For instance there was that case where somebody bought a blue
| check for a fake Eli Lilly account, tweeted "insulin is now
| free", allegedly caused the stock to drop, and the media erupts
| in outrage. Suddenly it is a big problem you can get a bogus
| blue check but, really, there is very little top stop me from
| buying some name like "elilil.ly", copying their template, and
| making a fake page with all the fake information I want. It
| would take their lawyers days if not weeks to get a preliminary
| injunction in a case like that, but truth is a site like that
| might go undiscovered for years. There is isn't a lot of talk
| about about the danger of website impersonation although there
| must be millions or billions scammed out of people yearly by
| fake web sites.
|
| Similarly there was the incident of J.K. Rowling getting
| dogpiled on Twitter that was only remarkable because (1) she
| was famous and (2) the story got picked up in the media. All
| kinds of people get dogpiled every day (often for being gender-
| nonconforming) and (as awful as it is) it's usually a free in
| the forest that falls without making a sound.
|
| In both of these cases only a few very people experienced them
| on Twitter and the situation quickly blew over. The stories get
| told over and over and over again away from Twitter, people may
| still be reading about them 20 years from now but it won't be
| on Twitter.
|
| Cases like this happen on all kinds of social media but
| Twitter's got a special attraction to people in the media that
| makes it all the more so.
| jeremyjh wrote:
| > It's just different bluechecks now.
|
| No, the bluechecks now are simply much worse. Previously,
| status was a selector, which takes into account the preferences
| of many, many people. Now, the primary selector is being an
| Elmo stan with $8. I've little interest in hearing from these
| people.
| nancyhn wrote:
| Some prefer The Annointed (before), others prefer The People
| (now).
| timmytokyo wrote:
| Musk stans are not "The People".
| flangola7 wrote:
| Musk is barely human to begin with, like attracts like.
| rsynnott wrote:
| It's very clear from the piece that what he means by the
| timeline is the (old) algorithmic timeline.
|
| Though, note that, even if you use Twitter purely on the linear
| timeline, you are still exposed to the algorithmic timeline,
| because, well, where did you think the people you follow were
| seeing the stuff that they retweet?
|
| Even without this effect, it would have changed; a lot of
| people are _no longer using it_. My linear timeline is
| dramatically different to how it was a year ago, because so
| many people have just stopped tweeting (though, actually, now
| that I look at it, I had to scroll back about a hundred items
| on my linear timeline to find something that wasn't a retweet;
| presumably something is once again broken.)
|
| > In fact, Twitter pre-Musk had this. It's just different
| bluechecks now.
|
| As far as anyone knows, old-Twitter did not prioritise bluetick
| tweets (at least not directly, though the selection criteria
| for bluetickdom did mostly select for people who were getting a
| lot of visibility for other reasons); it certainly did not
| promote them to the top of the replies. Naughty Ol' Mr Car
| framed this as a new thing that he was doing (and the sea of
| irrelevant bluetick responses on anything popular is certainly
| a new phenomenon).
| debrice wrote:
| It's easy to forget that prior to its acquisition, Twitter was a
| financial catastrophe. It still could be, but Twitter seems to be
| working hard at becoming a profitable business. Does praising its
| unsustainable past and lost designs makes us some kind of spoiled
| brats?
| woodruffw wrote:
| Why are we reconning this? Twitter was net positive for most
| quarters between 2018 and 2022[1]. They were never a money
| printing machine like Facebook or Google, but that wasn't their
| segment anyways.
|
| The phrase is "work smart, not hard." I'm not convinced that
| Twitter's leadership is doing either, and the market for
| Twitter advertising appears to agree.
|
| [1]: https://www.statista.com/statistics/299119/twitter-net-
| incom...
| ianburrell wrote:
| 2020 was disaster but that was true for lots of companies.
| 2021 would have been profitable except for settlement. In
| 2022, they started losing advertisers after offer but before
| sale. With moderate layoffs like other companies did, they
| would have been fine in 2023.
| optimalsolver wrote:
| *retconning
| idlewords wrote:
| Not to mention that they were clearly overstaffed,
| particularly at a time when there were almost no changes to
| the core product. All they needed to do to be extremely
| profitable was cut a couple of thousand employees and find a
| way to let power users pay for the service. In other words, a
| non-clown-car version of what Musk did later.
| zenkat wrote:
| Interesting perspective, considering ad revenue is down nearly
| 60% y/y.
|
| If they are "working hard at becoming profitable", it doesn't
| seem to be working.
| gilbetron wrote:
| All this reminds me of World of Warcraft. In 2004-2008ish, WoW
| was a cultural phenomenon. South Park made an episode about it,
| there was a famous question on Jeopardy about it, parents played
| it, celebrities played it. Then it started to fade away for many
| reasons. But as it was fading, people were asking what was the
| next MMO, what was the "WoW-killer". It never came - sure there
| were and are other MMOs (ESO, FF, GW2, etc) that people played,
| but nothing ever peaked like WoW, because WoW happened at a
| unique moment in time. It was right before social media exploded,
| and so people were able to be social with each other in the game,
| creating a unique community situation. All the other MMOs
| (mostly) grew up with social media, so the social effects were
| diminished. They were just games.
|
| I see an analog with Twitter. There will never be a Twitter
| killer, a Twitter 2.0. There will be successful, Twitter-like
| things (Threads, Bluesky) but nothing can be like Twitter because
| the era is gone. (Which is the point of essay, more or less).
| Something else will happen that will engross us for a while, and
| I'm curious what it will be (and really wish I could predict
| it!), but nobody, not even Twitter (just like WoW itself was
| never a WoW killer), will be another Twitter.
| kyleyeats wrote:
| Adam Devine was on Theo Von yesterday talking about pitching
| movies. He was bothered by executives who asked "Why should we
| make this movie now?"
|
| I was too until I read your comment. Devine just wanted to make
| a funny movie, but it's not enough to just be a good
| categorical example of a thing.
| the-printer wrote:
| > In the world of Almost Famous, Instagram would be the social
| network for the Stillwaters, the Russell Hammonds, the Penny
| Lanes. Beautiful people, cool people. Twitter was for the uncool,
| the geeks, the wonks, the wits, the misfits.
|
| Dichotomies like this are becoming trite in my opinion, even more
| so because this representation of "cool/not cool" doesn't ring
| true to my experience on either platform. Who/what falls in line
| with either category is largely subjective to the subcultures on
| the platform itself, among other things1.
|
| > Twitter was the closest we've come to bottling oral culture in
| written form.
|
| This is nice.
|
| 1. With the exception of Tumblr, in my experience at least,
| Tumblr seemed (entertainingly) hostile to the kind of people who
| probably went on to launch glowing Instagram personas.
| TeaDrunk wrote:
| Tumblr had their own sense of influence and a very different
| kind of person that would launch careers. IMO Tumblr was way
| more into thought pieces, deep fandom rants were encouraged,
| humorous quips that would go viral, and collages of
| pictures/aesthetics.
| gary_0 wrote:
| > The changes to the Twitter algorithm bulldozed over a decade's
| worth of Chesterton fences in a few months.
|
| Keeping the Chesterton fences in an established organization was
| something Musk was particularly ill-suited for. While some
| ossified companies can benefit from new management, things will
| only get worse if they throw the baby out with the bathwater.
| Musk's successes started out as greenfield projects that
| benefited from ignoring how the incumbents did things, but those
| days are over for him.
|
| Another thing Musk was ill-suited for was being the guy that did
| the mass layoffs. In the earlier days of Tesla and SpaceX, most
| talented engineers would at least be intrigued by a job offer
| from Musk. He had a knack for finding brilliant people who could
| make his mad schemes a reality. But now he seems to embrace the
| darker side of his cavalier persona (while also flirting with
| demagoguery), which has alienated a huge portion of the talent
| pool that all of his companies rely on.
| KnobbleMcKnees wrote:
| _Twitter hit upon some narrow product-market fit despite itself.
| It has never seemed to understand why it worked for some people
| or what it wanted to be, and how those two were related, if at
| all. But in a twist of fate that is often more of a factor in
| finding product-market fit than most like to admit_
|
| Hear, hear!
|
| Most internet companies have no deep or useful understanding of
| how they pioneered their early successes, or how to tackle the
| widespread growth decline that they are now encountering. Reddit
| is another fine example of this effect.
|
| I think the one thing Musk achieved in his _folie a un_ is
| putting Twitter on the observation deck for long enough that this
| fact became painfully obvious to many.
|
| I think what we're learning - in fact, what we already knew in
| the days of newsgroups, and IRC - is "if you build it, they will
| come". And 'it' can be pretty much any medium that removes the
| potential inconveniences of connecting with other human beings
| and voicing opinions to them.
|
| There really is no product beyond the people and the content.
| Google did well to realise this and turn itself into a huge
| parasite pretty much from day one: search aggregated the product,
| analytics kept tabs on what it was up to, and ads exploited both
| of the former.
|
| That's why we will continue to see VC money and modern social
| media companies throw themselves at the wall like spaghetti and,
| eventually for most, off the cliff like lemmings.
|
| The current fracture in social media may be the best thing that's
| ever happened to the internet ( _a few pennies in the hyperbole
| jar, sure_ ). Look how much competition we have now in the 'let
| humans get their opinions and memes to other humans' space:
| Mastodon, BlueSky, Lemmy, Squabbles, and so on.
|
| Sadly capitalism will once again inevitably demand a winner and
| we'll be back at whatever Twitter/Reddit 2.0 is but at least
| there's always hope for the cycle to begin again.
| PaulHoule wrote:
| Yep, the genius of Facebook was that it got ahead of the
| decline cycle. It's survived several half-lives of decay
| because it saw what happened to MySpace and they weren't going
| to let that happen. Getting a two-sided market for advertisers
| and consumers was a big part (if it is money you care about you
| want the effective ad network not a social media product that
| has other good attributes). Other parts have been continuous
| investment, buying lots of competitors, side bets like "The
| Metaverse", ...
|
| They understand completely that if they aren't moving forward
| they'll quickly wind up like Kodak, Polaroid or AMC. Contrast
| that to the average communications app (App X) which starts out
| with an onboarding process that works (like it really does find
| your microphone) but a few years later they get complacent and
| it "just doesn't work" despite having 1000+ engineers and an
| engineering blog about how badass they are. Now a friend tells
| you to try App Y and you'll say "I'm wary of this, look how App
| X sucks now" and they'll say "No, App Y is great, remember how
| App X was 5 years ago?". 5 years later App Z will take the
| place of Y and Y will take the place of X.".
|
| Most people don't get that this is the reason why communication
| apps are a technological backwater like narrowbody airliners,
| but Mark Zuckerberg does.
| gonzo41 wrote:
| I don't think Zuck is that clever or ahead of the curve. He
| built a giant social network that had guard rails, ie fill in
| the profile, rather than build the website (myspace). And as
| a consequence has ridden first mover advantage for the last
| 20 years.
|
| If Facebook hadn't been able to acquire Instagram or WhatsApp
| the internet of today would look totally different.
|
| Facebook got lucky because regulators in the US were asleep
| at the wheel.
|
| Anyway, facebook is a dark forest. Threads may have 70M new
| users but it's got instagrams TOS (no fun) and it's not
| prioritizing news. Journalists will come quickly and then
| leave because content will happen elsewhere.
|
| Maybe twitter will outlive musk. Maybe.
| PaulHoule wrote:
| I'd describe the acquisitions as a strategy to keep on top,
| particularly when they have the advertising machine to turn
| audiences into gold.
|
| What I don't understand is how funding for new social media
| startups has collapsed so completely in the U.S. Insofar as
| Facebook would buy (or want to buy) any potential
| competitor that seems like a pretty good exit plan. Today
| there is more concern that the government would get in the
| way but I think the funding collapsed happened before anti-
| trust got fashionable again.
|
| I think you're right that Threads won't appeal to
| journalists like the way Mastodon is not appealing to
| journalists but particularly Facebook deciding news is not
| great for business since news organizations are demanding
| to get paid and news is provocative and inflammatory (who
| wants to put an ad next to a video of a little bomb
| dropping out of a drone on someones head or next to
| somebody getting viciously dogpiled?)
| Terr_ wrote:
| Going a step further: It isn't just that some companies don't
| _know_ how they became successful, but that their success is
| largely up to un-repeatable factors we can only call "luck".
|
| Disregarding the influence of chance is a systemic issue,
| because humans are hard-coded to look for distinct reasons--
| stories--even when they are ridiculous. Consider how common it
| is for people who are flipping a coin to end up talking about
| "hot streaks" or the opposite outcome being "overdue."
| m3kw9 wrote:
| This guy wrote almost a book on how twitter "blew up", or you can
| go and use it and see if it is on that scale.
|
| Meta sensed a sentiment shift against twitter based on a some
| recent questionable moves by Elon, not a blow up.
|
| It's a good opportunistic move by Meta.
| LeafItAlone wrote:
| > or you can go and use it and see if it is on that scale
|
| I'll never know because I don't have a Twitter account and I
| can't seem to see Tweets without logging in anymore.
|
| Can you tell me what it's like now?
| PaulDavisThe1st wrote:
| Pretty much like it has been for the the last 5-10 years.
___________________________________________________________________
(page generated 2023-07-08 23:01 UTC)